Meeting Banner
Abstract #2813

Denoising Diffusion-Weighted Images by Using Higher-Order Singular Value Decomposition

Xinyuan Zhang 1 , Man Xu 1 , Zhe Zhang 2 , Hua Guo 2 , Fan Lam 3 , Zhipei Liang 3 , Qianjin Feng 1 , Wufan Chen 1 , and Yanqiu Feng 1

1 Biomedical Engineering, Guangdong Provincial Key Laborary of Medical Image Processing, Southern Medical University, Guangzhou, Guangdong, China, 2 Biomedical Engineering, Center for Biomedical Imaging Research,Tsinghua University, Beijing, Beijing, China, 3 Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States

Diffusion-weighted (DW) magnetic resonance imaging is widely used in clinic and research because of its ability to characterize the diffusion of water molecules within tissue. However, the DW images are usually affected by severe noise especially at high resolution and high b values, and the low signal-to-noise ratio may degrade the reliability of the subsequent quantitative analysis. Recently, a patch-based higher-order singular value decomposition (HOSVD) method was proposed to denoise MR images and demonstrated to outperform the well-known BM4D method. Compared with the conventional T1-, T2- and proton density (PD)-weighted images, DW images may contain more redundant information because that they are usually highly correlated across different diffusion directions. In this work, we proposed to simultaneously exploit the redundant information along diffusion directions and across spatial domain by using HOSVD in denoising DW images.

This abstract and the presentation materials are available to members only; a login is required.

Join Here